Summer Cruising aka Long Service Leave

The warm weather is winding down here but I did get a chance to get to the beach a couple of times. The water was a little brisk but on the refreshing end rather than numb appendages end of the scale.

I didn't get there as much as one would like but then our summer was a bit of a fizzer, I think we only had a couple of days over 40 and not many more over 35. After a dry winter we've had a drier summer so the garden needed constant watering, the veggies were a bit shit but i should be getting a decent crop of chillies - the only problem is they're about 2 months too late and will shortly stop growing but maybe some of the ones in the pots will keep going for a bit longer if i can find a sunnier spot in the garden.

Apart from that I'm just trying to enjoy my `long service leave' of indeterminable length. I've had some jobs suggested to me but it just turns my mind off, I feel really burnt out still. I've been trying to get into various hobbies from photography to video games to electronics, cooking, wood/metal work, sewing, reading ... but I just can't seem to get past a littler tinkering before losing interest in most things. I guess I need some specific projects, I should finish my workbench for one but it's a bit tricky getting shit without a car.

Trying to keep active socially as well, spending too much time at pubs and bars (which is getting a bit bloody expensive these days) but also trying to visit friends more often. Cycling a lot more than I have done in a long time although far short of where I've been in the past. Walking quite a lot. I lost a little weight and gained a lot of fitness at least. My broken hip hurts pretty much constantly but it isn't enough to bother me most of the time and whilst being active brings different pains it's better than sitting for lengthy periods so that's another factor driving me off my arse and potentially out of the house.

My "new" computer isn't doing much - it's just a web box most of the time. I keep the kernel and Slackware-current updated and have been keeping a loose eye on technology like Java but it's more out of habit than anything. I'm not doing any hacking - i'm still using playerz every day to play music but I was half-way through some work on playlists when I stopped. I guess the weather's been too good a lot of the time to waste it inside anyway.

Quit my job!

Decided to finally pull the pin on work. It's been coming for years but some issues regarding long service leave and some minor conflicts with the new manager finally tipped me over.

It's uncanny just how much my mood has lifted since the decision to leave - and is still up there a week later after a bit of a hard slog working on handover material.

I have absolutely no plans! I've got savings enough to tide me over for an extended period so I don't even have to rush to look for work.

Not sure what else i'll get up to but i'm sure along the way i'll have time to dabble in more Free Software than I have for a while and maybe start interacting with the wider "commnity" more. The music player continues to come along too and more on that later.

playerz: a music player

I got sick of mocp crashing every couple of days so i've finally got the music player I was working on last year into a state good enough to replace what I was using it for. It's still early alpha software and I don't have a project page for it or any documentation but the code is at https://code.zedzone.au/. It requires an old version of libeze, possibly the earliest that was checked in but I haven't checked.

Originally I was coding this for a beagleboard or the like where the player would be used without a graphical interface. It would automatically detect and scan new media and remember old media so re-inserting media would be fast. I also experimented with a 'talking' menu driven by a remote (Mele airmouse). Since i've now moved my old pc to under the TV i'm just targeting that now.

It consists of a bunch of services which comminucate using posix message queues.

music-player

Plays the music and allows navigation forward/reverse through the (shuffled) playlist. A rudimentary 'talking menu' has been implemented using espeak-ng but seems broken at the moment.

ffmpeg is used to parse and decode the media and to perform automatic volume leveling, and ALSA is used for audio output.

disk-indexer

Recursively scans mounted disks (or directories) looking for music files and gathers metadata from them and writes them to an index. It handles re-scans efficiently by only scanning for changes. It also generates the shuffle table across all files. lmdb is used for the database backend.

input-monitor

Monitors a raw input device for user input and sends commands off to the player. It is hard-coded to the Mele air-mouse but due to some issues with the USB stack device naming in Linux it can't always find the mouse device which is used for some of the controls.

disk-monitor

Checks for removable media to become available or be removed via raw kernel UEVENTs. New media is mounted read-only and passed to the disk indexer for a (re)scan. Removed media is marked as unavailable.

audio-cmd

A command line tool to send various commands to the player such as next or previous and so on.

mocp was having trouble with a collection of about 160K music files I have, and the TUI was pretty slow every time I started it. Indexing same said files takes about 20 minutes with playerz (I dunno roughly on par with mocp?) but restarting the player is basically instant.

I'm doing some web stuff for work at the moment so will probably look at some web control stuff, although for now the Mele air-mouse does most of what I need (skip to next song).

Even though i've done fuck all this year i'm getting a bit sick of work. Nothing interesting to work on and the company I work for has been trying to bilk me out of long service leave (13 weeks paid time off) - something I should have had 5 years ago. Really just want to retire but i'm a bit too young yet for that even if I have really nothing to work for anymore.

Going nowhere

Still going though.

At some point a few months ago I almost completely lost interest in coding - my new computer has just been collecting dust. This is the first time i've turned it on in a couple of months, and all I did at that time was upgrade the kernel. Same with this time actually but I also upgraded netbeans to 12.0 because I wasn't doing anything else.

Work is super-slow, partly due to COVID, partly due to timing with respect to breaking my leg, partly the project i'm on and a good dose of I don't give a fuck.

I've been spending a lot of time alone at home, again partly due to COVID but also plenty of I don't give a fuck. Even on a cold day if there's sun about sitting in the backyard with a book and some (nettle) tea is preferrable to drinking beer alone in a cold pub. I seem to feel the cold terribly these days. I'm getting out when I feel the need to, the feeling comes and goes.

I've spent a bit of time playing games, mostly No Man's Sky and a lot of Dreams. Once I got to >$U 4billion in NMS there doesn't seem much point since I avoid the multiplayer but I still go poke around every now and then to see what's cooking.

Dreams is a lot more interesting so i've spent a lot of time in that. I worked on a few starts of games, but I find desiging levels isn boring as fuck so I usually don't get past working out the base gameplay and a demo level. Dreams logic is very limited so it makes doing much with it a bit difficult, but just the challenge of those restrictions is entertaining enough, at least in measured doses. There's some really impressive stuff in there though if you can find it. Should be interesting to see what the PS5 can do with it, and if there's a new PSVR coming.

Till next time ...

New PC Gets a Box

I finally built a box for the computer I bought back in December.

I made it out of 12mm MDF. MDF really sucks, it's heavy and it's fragile but its' cheap and available - here in Adelaide you can't get much (for any decent price) and I don't have the machinery to dress wood anyway.

Because you need to take the lid off to get to the insides you can't just screw into the MDF directly, partly with 12mm board. So you need to get some way of strengthening the wood.

The only part I could find locally was an M6 screw insert but it requires a 9mm hole so it's too wide for the 12mm MDF. So I had to create wooden lugs to mount these. I was going to try using MDF but it just seemed like too much risk with such shitty wood and it would make drilling the holes more difficult and require more days for glue to set and so on. I used some salvaged scrap and mounted them using 2 screws to the MDF (and glue).

I started with a 450mmx1800mm piece and used under half - I got my brother to buy it when I was on crutches and I hadn't worked out a design yet. I cut 2x280mm pieces, which I then cut a 150mm strip from each - these became the sides and top and bottom. Then I cut one more 174mm strip which became the front with a small piece of scrap. Basically the sides go the full length of the case, and front end-caps everything.

So I didn't take any photos until I had the box nearly done so I'll briefly describe the steps. First I cut all the pieces and sanded down the edges. In particular the top and bottom pieces had to be true and square on at least 3 sides. I worked out where I was going to mount the top fans and cut the holes for those using a jigsaw, a compass helps to centre the circle properly.

I don't have any corner clamps so instead i used a square and a square piece of wood (the 'scrap' from the cutting) to align one piece at a time, glue and screw, clamp, and leave overnight to set. I did that for the bottom first. Then I did it with the top, also ensuring it was parallel to the bottom and square to the front. And finally the front could just be screwed and glued on. Fool: I didn't roughen up the flat of the MDF enough so the glue isn't as strong as it should be - after all the painting and sanding the rear top began to split off but only a few mm worth (I'd kept the screws out while I was painting).

I had intended to use some dowels, I even bought an overpriced dowel jig for the purpose - but I bought the wrong sized dowels (I think they were on the wrong hook in the shop and I didn't check) and the dowel jig was absolute junk. I got one OK hole and then it wouldn't sit in the right place. Helped by the fucking shitty 'dowel drills' which I also bought which just weren't straight. $130 down the drain basically. Fuck Bunnings. Why does everything they sell have to be fucking junk.

Once I had a box to work with I set about solving how to mount the side panel. I cut a strip from a bit of old wood and then made lugs out of them. For the threaded insert I did an 8.5mm hole deep and a 9mm hole half way for an extra-snug fit. It was tricky screwing them in straight. Idiot: I should've spent more time squaring up the strips! Luckily it's good enough.

I decided to go with screws to mount the lugs, which I countersunk. The screws are just at the limit of fitting within the 12mm MDF but they seem OK. Later on I glued them in place as well.

I drilled the holes in the front panel. First a couple by measurements then I made a pin to drop into the inserts which marked the spot. I should've got some dowel markers - should've just got those instead of the shitty jig in the first place. Mistakes were made, not the last. I countersunk them to match the stainless steel bolts and then screwed it all together.

Then took a couple of afternoons to sand it all square.

Next was positioning and mounting. The hard constraint was at the top - where the fans go. At the bottom the GPU and the lugs collide so that set another hard limit, and the PSU just had go to where it fit. I think I should've positioned the mounting board a bit further toward the bottom (right) but it's no big deal and something that isn't too difficult to change.

The base-plate was from an old PC I took apart years ago to build a small case (which I never built), I'd chopped it up for an ITX board with a single slot which just happened to be almost exactly right but I did trim it a bit further to give more options for the PSU. I just screwed it to the wood with tiny screws.

The PSU was more work, I went through a couple of ideas but settled on a pair of brackets. A small angle on the bottom and a corner bracket for the top, leaving room for cables, airflow and any front-panel stuff behind it. The power cord was quite a problem, I had to hack away most of the cable support plastic and even the front so the IEC plug pushed in far enough and the wires didn't stick into the fan. At the time the modular power cables from the output were pressing up against the GPU so I couldn't move the PSU any further away from the fans - but with some adjustments I made more room. So the PSU could potentially be moved but I haven't done so and the IEC cable would still need some hacking anyway.

It took me a while to work out how to create support for the PCI bracket. A flat bit of metal? Some angle? The problem is that the IO panel and the GPU and the rest of the design mean I can't just use a piece of wood as the back plate as there isn't enough support area or there's no way to then get the computer together. And that means any PCI support wont be well supported either. I toyed with using the PCI mounting plate from the computer I'd chopped up but I couldn't quite work out a way to mount it (actually in hindsight this is one reason the base plate is where it is, it was just enough for 4 slots with the board against the base). That probably would've worked but instead I went with using a piece of square tubing. I cut a slot out at the bottom of the tubing so that it slides onto the end of the PCI bracket, and then I use another pair of lugs to mount the bar to the case.

It's a bit tight and probably slightly out of position but it's solid and supports the card well.

So I put it all together and started doing some heat testing. I made some vents in the side panel based on the position of the GPU and other constraints like the mounting lugs. The upper vents align with the GPU vents on this particular GPU.

The idea is the bottom ones feed cold air to everything and the higher ones can vent some GPU exhaust if they get hot. In practice air just seems to come from all of them. My GPU fan never goes above the bottom rate anyway, I don't play games so it isn't doing much more than running the desktop now.

I normally run eco mode (65W) on the Ryzen 3900X, but I turned that off to try it out. Running blender and the Mandelbrot OpenCL stuff from the previous blog the CPU got up to about 81C and the GPU in the 70s. There was a bit of a hot-spot near the front end of the GPU but it didn't seem too bad.

On the other hand the cables were stuffed in pretty badly, particularly behind the PSU and up against the GPU. So I worked out a way to fold them and tie them up and it improved airflow quite a bit. I also chopped up the HDD cable from the PSU and soldered on one of the fans - this motherboard only has a single fan header, and I got the PWM fans. It works better if they're both on 'flat out' anyway - they're only low rpm fans and it keeps the CPU cooler so it makes less noise. There's a shot later with this one.

Painting. It sucks. You paint, you clean, you sand, you paint, you clean, you sand. Makes a mess. Stinks. Expensive.

Actually before I painted it I detailed it.

• Added a 45° bevel to all the main edges. It's simple, helps protect the edges and I think it looks nice. I used a trimmer router for this.
• Added a smaller bevel to the case-side edges of the side panel and the corresponding edges on the main box. To protect the edges for a removable part but also to 'hide' any small misalignment.
• Cleaned up the fan holes - made them round and square.
• Cleaned up the vent holes.
• Created a hole in the front for the power switch.

And I fucked it up quite badly. 1x coat of sanding sealer, 3x white undercoats, and 3x enamel overcoats ... but the final overcoat is full of dust and hair and bubbles and isn't even enough. It could probably be fixed with a sand and recoat - but I just gave up.

I worked out later at least one mistake I made, shaking up the can rather than stirring it aerated the paint which came out when I used a roller. Oops.

Before I painted it I decided to cut matching slots on the other side panel, to help airflow a bit more but mostly for aesthetics. And of course I totally fucked it up - I put the upper slots in the wrong place because I measured from the inside and not the outside. Fucking idiot! That really pissed me off for days but what can you do eh? The base plate was always going to cover some of the upper rear slot but I could've cut some of it away, but I just gave up.

Here I've mounted the power button as well. I don't know what to do for the button - it's a bit fiddly making anything that will push onto it, but it works as a 'hidden push-button' hole in the case anyway.

I decided I should probably cover the fans, don't want a stray screw falling inside or something. I used some aluminium fly wire.

Nearly finished, fans, motherboard, power supply. Here the cables have been tidied up. The reset switch is just hanging out the back for now.

And that big-arse GPU is in now too! After tidying up the cables there is ample room.

I also detailed the galvanised square tubing - I ran a wire brush over it with the drill to give it a satin finish and then put on a couple of coats of Penetrol. I haven't used it before but it's supposed to be 'the shit' for a natural metal finish (I got it for my kilt belt buckle but haven't used it yet). The steel bar doubles as a mounting point for the magnetic antenna that came with the motherboard.

Due to the issues mentioned earlier there's no rear panel yet. Maybe i'll do some steel and attach to the bar or something. In hindesight the case could've been another 15mm long so the bar could be recessed ... but it's too late now (unless I get a shorter GPU). The case could have been any depth!

Despite 24 hours drying in good weather the Penetrol and paint were still a little tacky feeling (or rather, set but soft). Because of this I only did up the screws lightly, but perhaps I should've waited a few more days for the paints to harden more. I also should have sanded down the internal surfaces where the paint created a thicker edge too, already a couple of bits have pulled off but they're inside.

That's it!

It is still on the dining table because my computer desk is occupied with junk, my old Kaveri machine, and a giant old full-tower from work.

ZCL and Project Panama

So I've been really busy working on zcl again, porting it to the foreign-abi branch of project panama.

Well that's after I ruffled some feathers on the mailing list, I guess I just have strong opinions on C, but I thought I'd better put my money where my mouth was before upsetting anyone too much. To be honest the Oracle engineers have been the most polite and overwhelmingly patient of all the projects I've interacted with lately - guix guys just weren't very polite (that's not entirely fair, some were nice), the google guys for clspv were just condescending, and well the ffmpeg devs never did reply to my patch for a kinect indev.

Anyway back to the topic at hand. I wrote a fairly rushed but quite detailed overview of the issues I came across and some of the features I implemented or needed to change. I titled it JNI to Java with project panama. The prose doesn't flow very well but at least I ran it through ispell which is a habit I'm trying to get into.

I was originally just going to put into a readme in project panamaz but since I got so far with the development I just published it separately and created a new branch foreign-abi in the zcl repository. The article above has some checkout details about checking it out. It's still work in progress and I'm still making some internal changes but I added a fun little Mandelbrot zoomer to it. The javadocs are pretty shit but for what it's worth I've uploaded those too and will try to keep them relatively up to date while I work on it.

I'm also using the project to experiment with the maven and junit, and now a demo-runner snippet for GNU Make.

It was sort of because I was working in the panamaz project but I wrote almost all of this new code in emacs rather than netbeans. Damn I knew it pissed me off but I didn't realise it wasn't writing Java that was the problem, netbeans keeps just fucking getting in the way when you're trying to write code and those interruptions constantly break you out of 'flow'. I mean the completion and the real-time syntax checking is pretty good but I don't know why the editor has to be such an annoying prick all the time.

Verified Maven Central from GNU Make

Well after the previous post I tried looking at the signature verification step. There isn't too much to it.

define maven_func=
bin/lib/$2-$3.jar:
        mkdir -p bin/lib
        wget -O $$@ $(CENTRAL)/$(subst .,/,$1)/$2/$3/$2-$3.jar
bin/lib/$2-$3.jar.asc: bin/lib/$2-$3.jar
        wget -O $$@ $(CENTRAL)/$(subst .,/,$1)/$2/$3/$2-$3.jar.asc
        gpg --batch --verify $$@ $$< || ( rm $$@ ; echo "GPG verification failed, you may need to import the public key." )
setup: bin/lib/$2-$3.jar.asc
endef

If it fails you need to import the keys using gpg manually. This might be a bit annoying but otherwise you may as well not bother.

gpg will pass the check if the key simply exists in your key store and ignores any trust setting, but that's just the way gpg works.

Maven Central from GNU Make

I had a look at jmh yesterday. It's all driven by maven so that's a bit of a pain but using the ant example I trivially converted it to use GNU make.

As expected, it's somewhat fewer lines of code to just use a generic build tool than either of the other tools which are specifically designed for Java projects. I will integrate it into java.make eventually, although as I don't use any projects that require it I haven't done that yet.

So for this prototyping example you just define a couple of variables. The interesting one just lists the group:artifact:version in a relatively simple/readable format.

CENTRAL=https://repo1.maven.org/maven2

CENTRAL_JARS=					\
 org.openjdk.jmh:jmh-core:1.18			\
 org.openjdk.jmh:jmh-generator-annprocess:1.18	\
 net.sf.jopt-simple:jopt-simple:4.6		\
 org.apache.commons:commons-math3:3.2

And well that's about it. It hooks into the makefile system at the right place to download the libraries. I don't do any signature checks but I can't imagine that would be very to add either.

It only requires a simple macro to download the packages.

define maven_func=
bin/lib/$2-$3.jar:
        mkdir -p bin/lib
        wget -O $$@ $(CENTRAL)/$(subst .,/,$1)/$2/$3/$2-$3.jar
setup: bin/lib/$2-$3.jar
endef
.PHONY: setup
$(foreach nver, $(CENTRAL_JARS), $(eval $(call maven_func,$(word 1,$(subst :, ,$(nver))),$(word 2,$(subst :, ,$(nver))),$(word 3,$(subst :, ,$(nver))))))

Although calling it is a little clumsy, but that's a hidden detail.

To hook it in you have to add jmh-core and jmh-generator-annprocess to the class path of the javac invocation.

bin/.classes: setup $(SOURCES)
        javac -d bin/classes \
                 -cp bin/lib/jmh-core-1.18.jar:bin/lib/jmh-generator-annprocess-1.18.jar \
                $(SOURCES)
        touch bin/.classes

jmh is typically run using a standalone jar file but for simplicity it can just be run in-place from the build directory. Creating a standalone jar wouldn't be that much hard to add, just a few invocations of jar to explode and repack the classes.

bench: bin/.classes
        java -cp bin/classes:bin/lib/jmh-core-1.18.jar:bin/lib/jopt-simple-4.6.jar:bin/lib/commons-math3-3.2.jar \
                org.openjdk.jmh.Main

In the final iteration I will add some code to the macro to create this classpath so it is easier to use.

panama/jextract

Regarding a previous post i've discovered that the annotation mechanism i've been using is probably not the approach to use and there is a lower-level interface. I'm still yet to start on supporting that in the code generator but i've nutted out most of the machinery it will require.

I've bugged the panama developers quite a bit about what I see as some shortcomings in the api. Their idea of a 'C api' is very limited, for example they think that Java shouldn't be able to dereference a C-allocated pointer without signifying 'unsafe' code and passing some scary arguments to the compiler and jvm.

Sure, many modern api's want to wrap trivial memory operations in multiple layers of 'accessor' functions but even those typically return a const char * for a string so you can't even access those without a mess. This seems a little odd because what you can do without resorting to 'unsafe' api's has no safety guarantees already, and infact you can trivially implement an arbitrary whole-memory-access function by just invoking memcpy. Anyway there is hope it is changed by release because it's an unecessary restriction that can be bypassed trivially and just adds more developer work for no good reason. In my experience such complexity just leads to more of the type of errors they're trying to safeguard against.